After excluding missing or incomplete SF36 surveys, 1210 participants remained.
After excluding missing or incomplete Neuro-QoL surveys, 1230 participants had filled out the survey
Or per number of illnesses:
Looking at nqol between top 10 most reported illnesses (and adjusting to Holmes-Bonferroni method for adjusting P-values to reduce false positives) However this allows for double dipping which may make illnesses seem closer related in presentation due to common participants reporting QoL for both
Due to how insanely interconnected some of these ADs are, concerned about double dipping. So, after assembling the df with top 10 ADs, grouped by record_id and removed duplicates. This left 556 entries and gives an idea how neuro-qol scores differ without overlapping illness. Unique entries per illness:
image
Definitely don’t have the numbers to look at change in nero-qol between diseases, by number of diseases.
Let’s have a look at that with SF36 as well:
image
image